You are on page 1of 13

Overview Articles

The Oxidative Cost of Reproduction:

Theoretical Questions and
Alternative Mechanisms

The role of reproduction in increasing oxidative stress may be fundamental to understanding aging and the evolutionary trade-off between
survival and fecundity. However, contradictory results among experimental studies have challenged the oxidative cost of reproduction. Limitations
in experimental design may explain inconsistent findings. Nonetheless, some authors have argued that the hypothesis is founded on the faulty
assumptions of an energy-based allocation trade-off and a direct positive link between metabolic rates and the production of reactive oxygen species
(ROS) by the mitochondria. We propose that evolutionary trade-offs do not require the allocation of limiting resources and that reproduction may
result in ROS-induced oxidative stress without increased mitochondrial ROS. We discuss the previously published oxidative-shielding hypothesis
and propose two new hypotheses: hormesis and extortion for reproduction. These hypotheses aim to explain the counterintuitive results indicating
that there is less oxidative damage in animals allowed to breed compared with those prevented from reproducing.

Keywords: hormesis, life history trade-offs, senescence, oxidative stress, terminal investment.

T he cost of reproduction is a cornerstone of and longevity (e.g., De Jong and van Noordwijk 1992). An
evolutionary theory (e.g., Zera and Harshman 2001).
Recently, oxidative stress has been proposed as a new com-
oxidative cost of reproduction would also support evolution-
ary theories linking reproduction and aging (Williams 1957,
ponent of that cost. Oxidative stress can be defined as the Kirkwood 1977). Along these lines, the disposable soma
imbalance between the rate of production of reactive species theory (Kirkwood 1977) suggests that the somatic line of the
(mainly reactive oxygen species; ROS) by cell metabolism and organism (the soma) is maintained (e.g., by antioxidant
the level or efficiency of antioxidant defenses (i.e., including mechanisms) while favoring reproduction (the activity of
repair mechanisms and intrinsic passive defensese.g., cell the germline). When this is not possible, the soma becomes
membrane composition)which leads to damage to lipids, disposable, and death occurs. The oxidative cost of repro-
proteins, or DNA(Halliwell and Gutteridge 2007). We must duction would magnify this effect, accelerating aging and
also consider regulatory processes controlling ROS genera- promoting a trade-off in the allocation of resources (i.e.,
tion and the disruption of redox-signaling mechanisms as energy; see below) between reproduction and survival.
components of oxidative stress (Jones 2006). Redox signal- In the early 2000s, two studies demonstrated that the short-
ing is the specific, usually reversible, oxidation or reduction term (a few hours) survival of fruit flies exposed to paraquat
modification of cellular signaling pathway components by a (an ROS generator) declines when animals are stimulated to
reactive species or by the shift in redox state of a responsive invest in reproduction (Salmon etal. 2001, Wang etal. 2001).
group such as a dithioldisulphide couple (Forman et al. These were the first studies explicitly testing whether repro-
2004). Redox-signaling disruption has often been linked to duction leads to oxidative stress and whether oxidative stress
pathological states (e.g., Go and Jones 2011). induced by reproduction reduces longevity. This could, in
Oxidative stress has traditionally been associated with the turn, explain aging. This idea defines the oxidative cost of
aging process (Sies and Cadenas 1985, but see, e.g., Barja reproduction within the framework of two complementary
2013) but may potentially shape other life-history traits (e.g., hypotheses: (1) reproductive effort increases oxidative stress,
Isaksson et al. 2011). Therefore, the hypothesis that repro- and (2) oxidative stress due to reproductive effort reduces
duction leads to oxidative stress has attracted the attention of lifespan. Since these initial studies, most research has tested
evolutionary biologists, because it may provide a mechanism the first hypothesis, frequently reporting nonsignificant or
for the hypothesized negative link between reproduction even counterintuitive results (i.e., breeding animals may

BioScience XX: 113. The Author(s) 2016. Published by Oxford University Press on behalf of the American Institute of Biological Sciences. All rights
reserved. For Permissions, please e-mail:
doi:10.1093/biosci/biw176 Advance Access publication XX XXXX, XXXX XXXX XXXX / Vol. XX No. X BioScience 1

Overview Articles

endure lower oxidative stress than do nonbreeders; e.g., and Drosophila mutants that are able to live longer without
Speakman and Garratt 2014, Blount etal. 2016). The second apparent fecundity loss or that show a reduction in fecun-
hypothesis is, however, relevant from an evolutionary point dity without increases in lifespan (Flatt 2011 and references
of view. A cost can only be considered as such when it trans- therein). Nonetheless, we must note that these studies
lates into a lower contribution to the gene pool of the next mostly analyzed effects on egg production, with unknown
generation (i.e., Darwinian fitness). This can result from impacts on lifetime reproductive success, offspring quality
lower survival (reduced longevity), but also from reduced or fecundity. In fact, we should ask why these mutants are
reproductive success. In fact, the second hypothesis should apparently absent in the wild (Briga and Verhulst 2015).
be redefined to include the possibility of a reduction in the However, the cited network is intimately related to the
breeding output of subsequent reproductive events. metabolism of energy macromolecules (glucose and lipids;
The lack of consistent evidence supporting the first Hansen etal. 2013), which suggests that signals could play
hypothesis has been explained by some limitations in experi- a role in regulating trade-offs that will only be fully under-
mental design (Metcalfe and Monaghan 2013) but also on stood when the complete mechanism is revealed.
the basis of potential theoretical problems (Speakman and Social insects are also paradoxical for evolutionary theo-
Garratt 2014). In this article, we provide (a) counterargu- ries of aging. Queens are long lived compared with other
ments to the two main theoretical criticisms of the oxidative- castes but also invest more in reproduction, at least in terms
cost-of-reproduction hypothesis and (b) explanations for of egg production (workers also invest in reproduction via
the contradictory evidence suggesting that reproductive queen and colony maintenance). Seehuus and colleagues
effort may, at least in some cases, reduce oxidative stress. To (2006) proposed that the reproductive regulatory pathways
address the last point, we discuss several hypotheses. The were remodeled in the honey bee (Apis mellifera) to extend
recently proposed oxidative-shielding hypothesis (Blount life. In particular, these authors showed that vitellogenin, a
et al. 2016) and two new alternative or complementary protein involved in the production of egg proteins in many
hypotheses are explained and debated. To conclude, some taxa including vertebrates (e.g., Olsson et al. 2009), also
final remarks and directions for future research are provided. functions as an antioxidant prolonging the lifespan. This
implies that queen bees are able to simultaneously invest in
First theoretical problem: The requirement of an reproduction and survival. However, this mechanism seems
energy allocation trade-off to be restricted to eusocial Hymenoptera. A direct causal
The first theoretical problem argued to explain nonsig- link between vitellogenin and longevity has not been estab-
nificant results obtained when testing the oxidative cost of lished in vertebrates and is unclear in other invertebrates
reproduction is that there is no trade-off in the allocation of such as C. elegans and fruitflies, in which high vitellogenin
energy between the maintenance of a good oxidative status (Vg) expression is even linked to reduced longevity (e.g., Ren
(maintenance of antioxidant or repair mechanisms to attain and Hughes 2014, Seah et al 2016). Moreover, the genetic
low oxidative stress and survive) and reproduction (i.e., background of eusocial insects has led to the interpretation
Speakman and Garratt 2014). This idea was first formulated of the colony as a superorganism, with queens acting as the
as an answer to a previous debate emphasizing that experi- germline and workers as its disposable soma (e.g., Flatt etal.
ments manipulating breeding effort under captivity com- 2013). Therefore, testing the OCR would require the com-
monly involve ad libitum resources (e.g., food, water, and parison of different queens or hives on the basis of reproduc-
temperature; Metcalfe and Monaghan 2013). Under these tive output but not different castes within a hive.
supposedly optimal conditions, animals should be able to In any case, we propose that the requirement of an
simultaneously invest in reproduction and antioxidants or energy-based trade-off in the oxidative-cost-of-reproduc-
repair mechanisms, masking the trade-off. tion hypothesis is unnecessary considering two arguments:
The idea of an absence of an energy allocation trade-off in (1) nonenergetic resources (micronutrients essential for
the oxidative cost of reproduction was based on intriguing antioxidant protection, such as some vitamins, carotenoids,
findings obtained in fruit flies and Caenorhabditis elegans or cofactors in antioxidant enzymes) could be involved (see
(see Speakman and Garratt 2014). Animals with ablated also Mller et al. 2000, Alonso-Alvarez and Velando 2012,
germlines live longer, but those whose whole gonads were Blount etal. 2016), and (2) the trade-off can also be estab-
removed did not (Hsin and Kenyon 1999). A complex net- lished without any limiting resource at all.
work of at least four interconnected molecular signaling The second argument is probably the most definitive one.
pathways generated in somatic tissues, probably including Tatar and Carey (1995) proposed that reproduction is not
part of the gonad, promotes increased lifespan but is blocked an energetic resource sink limiting self-maintenance but
by signals from the germline (Hansen etal. 2013). This net- instead causes direct damage, impairing or inhibiting those
work suggests that biochemical signals could regulate lon- processes involved in self-maintenance (see figure 1; Tatar
gevity and reproduction simultaneously, independent of the 2001, Zera and Harshman 2001). Resources would not be
availability of energy resources (Edward and Chapman 2011, necessary for a trade-off if the associated cost is unavoidable
Alonso-Alvarez and Velando 2012). Moreover, reproduc- (i.e., an obligatory cost; Speakman 2008, Alonso-Alvarez
tion and longevity seem to be uncoupled in Caenorhabditis and Velando 2012). The same idea was suggested in our first

2 BioScience XXXX XXXX / Vol. XX No. X

Overview Articles

Reproducon metabolism, reducing ROS generation

(e.g., Speakman 2008, Salin etal. 2015).
OS The mechanism allows individuals with
high metabolic rates to reduce ROS pro-
duction (Speakman and Garratt 2014
and references therein).
Acquision Resources Allocaon O
Obligatory cost
Speakman and Garratt (2014) argued
that the apparent lack of a direct link
between metabolic rate and mitochon-
drial ROS production is a key weakness
O damage
of the oxidative-cost-of-reproduction
hypothesis. However, recently, some
authors have theoretically (in mathe-
Figure 1. The classical resource-allocation trade-off based on the Y model matical models) and empirically (in an
from de Jong and van Noordwijk (1992) and the obligatory trade-off based on experiment with insects) demonstrated
an inevitable cost (Tatar and Carey 1995). The investment in a trait can not that oxidative damage is more closely
only divert limiting resources for another trait (e.g., energy, macronutrients, related to the biosynthetic rate (which is
and antioxidants) but also trigger an inevitable cost, such as oxidative damage. involved in growth but also in reproduc-
A direct trade-off may thus coexist with a resource allocation trade-off. tion) than to metabolic rates (see Hou
However, this also means that the oxidative cost of reproduction may involve and Amunugama 2015, Amunugama
a trade-off without involving limiting resources. Abbreviation: ROS, reactive et al. 2016). These studies also suggest
oxygen species. that variation in the percent free radical
leak or uncoupling are less relevant. We
also argue that ROS generated from cel-
study on the oxidative cost of reproduction (Alonso-Alvarez lular activities other than respiration can potentially lead to
etal. 2004; see also Wiersma etal. 2004). We proposed that an oxidative cost of reproduction.
the higher metabolic rate during reproduction should inevi- Mitochondrial oxidative phosphorylation is not the only
tably induce higher mitochondrial ROS production and, endogenous source of ROS. Extramitochondrial enzymes
accordingly, oxidative stress. However, this scenario leads us are important producers (figure 2). Among them, the
to the second theoretical problem, which we discuss below. major contributors to ROS production, at least in mam-
malian models, are the nicotinamide adenine dinucleo-
Second theoretical problem: Metabolic rates are tide phosphate (NADPH) oxidases (also NOX), which are
uncorrelated to ROS production involved in superoxide production (Bedard and Krause
Reproductive animals commonly face higher metabolic rates 2007, Covarrubias et al. 2008). They are present in most
compared with nonreproductive ones (e.g., Speakman 2008). body parts, including the immune system (i.e., phagocytes)
As we mentioned above, it was originally proposed that this and reproductive tissues (Bedard and Krause 2007). In fact,
could lead to oxidative stress, explaining the oxidative cost NOX is the only enzyme whose main function is the produc-
of reproduction (Alonso-Alvarez etal. 2004, Wiersma etal. tion of ROS and is involved in redox signaling (Drummond
2004). This assumes that ROS are produced in direct pro- and Sobey 2014). It has been suggested that NOX isoforms
portion to metabolic rate as an unavoidable result of the may generate ROS in some tissues, even at a higher level
activity of the mitochondrial electron transport (respiratory) than mitochondria (e.g., Krause 2007, Brown and Griendling
chain, whose main function is to produce adenosine triphos- 2009, Dymkowska etal. 2015). Furthermore, we should con-
phate (ATP) by oxidative phosphorylation (e.g., Speakman sider that mitochondrial and extramitochondrial ROS might
and Garratt 2014). However, the link between metabolic impose different costs. Accordingly, in some Caenorhabditis
rate and oxidative stress has been seriously questioned. mutants, mild levels of mitochondrial superoxide extend the
Variation in the percent free radical leak (the percentage of lifespan (see mitohormesis below), whereas extramitochon-
total electron flow leading to ROS generation in the respi- drial superoxide may accelerate aging (Schaar etal. 2015).
ratory chain) causes the disproportionality between ROS Interestingly, high levels of energy macronutrients acti-
production and oxygen consumption (Barja 2013). During vate NADPH oxidases. High circulating glucose levels lead
mild aerobic exercise, a high rate of oxygen consumption to an increased ROS production by NOX in rats (Serpillon
leads to low local partial O2 pressure in the mitochondria. etal. 2009). Hypertriglyceridemic mice showed a high activ-
This, in turn, increases the efficiency of the respiratory ity of NOX and xanthine oxidases in the liver and higher
chain reducing ROS production (detailed in Barja 2013). hepatic levels of lipid peroxidation, protein carbonyls, and
Mitochondria also produce heat instead of ATP by uncou- oxidized glutathione (Alberici et al. 2009). In contrast,
pling oxidative phosphorylation. Uncoupling proteins in the the same mice showed elevated body metabolic rates and
inner membrane (UCPs) short-circuit mitochondrial energy mild mitochondrial uncoupling, leading to reduced ROS XXXX XXXX / Vol. XX No. X BioScience 3

Overview Articles

Figure 2. A simplified diagram illustrating the extramitochondrial sources of reactive oxygen species (ROS; based on
Bedard and Krause 2007, Covarrubias etal. 2008, Drummond and Sobey 2014). Nicotinamide adenine dinucleotide
phosphate (NADPH) oxidases are considered the main extramitochondrial ROS source within the cell (see the main text).
They have mostly been described in the cell membrane, where they release superoxide (O2-) to the extracellular space. They
can also be found in the nucleus and endoplasmic reticulum membranes. Xanthine oxidases (XO) are other important
sources of ROS. They are located in the cell membrane, peroxisome, and cytoplasm, releasing hydrogen peroxide from
hypoxanthine or xanthine (purine bases). Nitric oxide synthases (NOS) are located in the cytoplasm, releasing nitric oxide
(NO) or superoxide from L-arginine depending on its state (i.e., coupled or uncoupled NOS, respectively). ROS generated
by NADPH oxidases also elicit ROS formation from XO and NOS. Other ROS sources include lipoxygenases (LOX) and
cyclooxygenases (COX), which are located in the cell membrane and generate peroxides from fatty acids (FA). LOX can
also be found in the endoplasmic reticulum. In addition to XO, the peroxisome holds fatty acyl-CoA oxidases, generating
hydrogen peroxide. The cytochrome P450 in the endoplasmic reticulum also contributes to ROS production by releasing
superoxide. Finally, P450, COX, and even NADPH oxidases (NOX4; Sakellariou Jackson and Vasilaki 2014) are involved
in the respiratory chain in the mitochondrion.

production by mitochondria (Alberici et al. 2009). These key to activating extramitochondrial ROS sources. This may
authors interpreted these findings as a protective response ultimately promote oxidative stress among breeders as a part
against a rise in ROS levels due to extramitochondrial of the direct obligatory cost of reproduction (see the section
enzymes. We must note that the liver is an important storage on the extortion hypothesis below).
site for lipids in small mammals (which represent the animal
models in most studies testing the hypothesis; e.g., Metcalfe Hypotheses to explain lower oxidative damage
and Monaghan 2013, Blount etal. 2016), with its mass and among breeders
metabolism increasing during reproduction (Speakman We have presented two theoretical arguments that may
2008). Therefore, we can hypothesize that, rather than an explain the lack of significance in the results of many studies
overall increase in metabolic rates, the increased levels of testing the oxidative cost of reproduction, but why do breed-
macronutrients that allow for reproduction could be the ing animals surpass nonbreeding individuals in antioxidant

4 BioScience XXXX XXXX / Vol. XX No. X

Overview Articles

Table 1. Assumptions and predictions in hypotheses explaining lower oxidative damage in breeders compared with that
in nonbreeders.
Oxidative shielding Hormesis during reproduction Extortion for reproduction
Summary Breeders preemptively activate a Reproductive effort triggers a In nonbreeders, the maintenance of
decrease in oxidative damage to avoid compensatory response reducing large macronutrient amounts required for
future fitness reductions oxidative damage and improving reproduction increases oxidative damage,
fitness forcing animals to breed as early as
Assumptions A1 Some unknown physiological Oxidative stress or some unknown Some signaling pathway exists and is
mechanism exists and allows for redox signal activated by oxidative activated in nonbreeders to facilitate
synchronizing oxidative shielding with stress triggers the hormetic reproduction via behavioral changes
the start of reproductive effort mechanism

A2 Reproductive investment can High levels of macronutrients increase

constitute a mild stressor oxidative damage and/or alter redox signaling
A3 Reproductive opportunity can naturally
be constrained while experiencing high
macronutrient availability
Predictions P1 Breeders show lower oxidative-damage Breeders transitorily endure mild Nonbreeders increase oxidative
levels at the start of reproduction oxidative stress levels damage when circulating or storing high
compared with nonbreeders macronutrient levels
P2 Oxidative shielding remains activated Breeders subsequently lower Nonbreeders reduce oxidative damage when
as long as needed to protect offspring oxidative stress and show breeding, but damage values could still be
lower damage values than do higher than those in nonbreeders carrying
nonbreeders lower macronutrient amounts (Fig. 4)
P3 Oxidative shielding should mostly be The decline in oxidative damage The oxidative damage should first be
present in tissues influencing offspring should be present in any tissue present in tissues maintaining high
quality (i.e., reproductive tissues favoring adult fitness macronutrient levels
including gametes)
P4 The offspring increases quality and Parents enlarge lifespan and/or Nonbreeders maintaining high
survival prospects fecundity macronutrient levels reduce lifespan
via oxidative damage
P5 The mechanism becomes inefficient The mechanism becomes The mechanism should act or be reinforced
at older ages because of the inefficient at older ages because at advanced ages as future breeding
accumulation of oxidative damage of the accumulation of oxidative opportunities fade
P6 The artificial inhibition of ROS The artificial inhibition of ROS production
production or antioxidant or antioxidant supplementation delays
supplementation during reproduction in nonbreeders circulating
reproduction reduces hormesis or storing high macronutrient levels
and its fitness consequences
P7 Nonreproductive animals with large
macronutrient levels experience behavioral
changes to increase breeding opportunities

protection? This finding has been relatively frequent, par- of reproduction in females of homeothermic vertebrates
ticularly among studies manipulating reproductive status (Blount etal. 2016). These authors propose that reductions
(comparing breeders versus nonbreeders) in females of in oxidative damage are pre-emptive and function to shield
mammal species (reviewed in Blount etal. 2016), although mothers, and in particular their gametes and developing off-
examples in male mammals (e.g., Garratt et al. 2012, spring, from critical levels of oxidative insults that inevitably
Salomon etal. 2013) and birds (Costantini etal. 2014) also increase as a consequence of reproductive effort (Blount
exist. Should we accept that reproduction reduces oxidative etal. 2016, p. 495). We assume that, to be preemptive, some
stress, at least under some circumstances? Reduced levels of unknown physiological mechanism should synchronize
oxidative damage in breeders compared with those of non- the oxidative shielding with the start of the reproductive
breeders would potentially favor longevity, challenging evo- effort (table 1, A1). Moreover, the cited reductions could be
lutionary theory proposing the classical trade-off between due to lower ROS production and/or increased antioxidant
reproduction and survival (e.g., De Jong and van Noordwijk protection (Blount et al. 2016). The authors also proposed
1992). To explain these counterintuitive findings, three that increased UCP activity and estrogen-mediated signals
hypotheses are discussed (see also table 1, figures 3 and 4). favoring antioxidant protection may be involved (Blount
etal. 2016). The idea was mostly supported by mammalian
Oxidative shielding.This hypothesis was formulated on the studies showing lower oxidative-damage levels in breeding
basis of results from meta-analyses including both cor- females compared with those of nonbreeding ones (Blount
relational and experimental tests of the oxidative cost et al. 2016). The statistical effect was driven by decreased XXXX XXXX / Vol. XX No. X BioScience 5

Overview Articles

1. Nonreproducve stage 2. Prereproducve stage 3. Reproducon

Reproducon Reproducon

Pre-empve protecon (-)

Survival Survival Survival

1. Nonreproducve stage 2. Prereproducve stage 3. Reproducon

Reproducon Reproducon Reproducon


Survival Survival Survival


1. Nonreproducve stage 2. Prereproducve stage 3. Reproducon

Reproducon Reproducon Reproducon

(+) ROS
Oxidized compounds ( )
( ) ( )
Survival Survival Survival

Figure 3. Hypotheses aiming to explain the lower oxidative damage among breeders compared with that among
nonbreeders. (a) The oxidative-shielding hypothesis implies that animals preemptively (stage 2) reduce reactive oxygen
species (ROS) production and/or increase antioxidant levels in order to protect themselves and their offspring from
oxidative damage during reproduction. The mechanism might be activated in advance of the reproductive event. (b) The
hormesis-during-reproduction hypothesis suggests that certain intermediate ROS or oxidative-damage levels generated
by reproduction trigger some unknown redox-signaling pathway activating a hormetic (compensatory) response that
should subsequently reduce oxidative damage. The mechanism would be initiated from the start of the reproductive
event. (c) The extortion-for-reproduction hypothesis assumes that animals increase body stores and/or circulating levels
of macronutrients (MN; i.e., carbohydrates, lipids, and proteins) in order to reproduce (1, 2, 3). Macronutrients not
diverted to reproduction would become oxidized or stimulate NADPH oxidases, increasing ROS and inflicting oxidative
damage (2). This process implies an obligatory cost in the trade-off between reproduction and self-maintenance (figure
1). High ROS or oxidative-damage levels would trigger redox signals encouraging reproduction as early as possible.
When reproduction takes place (3), macronutrient amounts decline; therefore, the oxidative damage is reduced. At
that time, ROS, perhaps derived from increased enzyme metabolism in cells (figure 2), would drive the oxidative cost of
reproduction, but circulating oxidized compounds due to resource allocation to reproduction may still contribute to that
cost. We predict that the efficiency of (a) or (b) mechanisms becomes weaker with time because of aging, whereas (c) should
be reinforced with age as opportunities for breeding decline (see also table 1).

6 BioScience XXXX XXXX / Vol. XX No. X

Overview Articles

levels in organs to remain low? Hormone-

Nonreproductive Prereproductive Reproduction related mechanisms (estrogens) or mito-
stage stage chondrial uncoupling are not stable
processes. Perhaps we should exclusively
predict a preemptive increase in protec-
tive mechanisms (via antioxidant levels
or reduced ROS production) and then a
Extortion subsequent reduction in oxidative dam-
Oxidative stress

age (e.g., figure 3a). Second, although

Hormesis reduced levels of oxidative damage in
follicles may support the protection of
the offspring, why should we expect
the same shielding process in somatic
Oxidative shielding
tissues, such as the heart or liver (table
1, P3)? Third, why would somatic non-
reproductive tissues in breeders show
lower (not merely equal) damage levels
than those in nonbreeders? Finally, the
hypothesis is difficult to distinguish from
a hormetic mechanism (below).
Time On the other hand, recent correla-
Figure 4. The proposed dynamics of oxidative-damage values throughout tional evidence from the same research
nonreproductive and reproductive periods in the life of an animal. The team (Vitikainen et al. 2016) seems to
extortion-for-reproduction hypothesis (dotted line) predicts an increase in support the hypothesis. Wild female
the oxidative damage associated with the accumulation of macronutrients banded mongooses (Mungos mungo)
in blood and/or storage tissues. The oxidative-shielding hypothesis (dashed showed lower plasma levels of oxida-
line) predicts a preemptive decline in oxidative damage, which should imply a tive damage in lipids (malondialde-
decrease in advance of reproduction effort. This could be triggered by hormones hyde, MDA, levels) during pregnancy,
(e.g., estrogens; see main text) and could start abruptly. The duration could in contrast to other mammalian species
subsequently vary depending on reproductive strategies (e.g., mammals could (below). Moreover, contrary to previous
shield pups from the embryo stage until weaning, whereas birds or reptiles findings in other species (Blount et al.
should do so only until oviposition). Finally, the hormesis-during-reproduction 2016), Vitikainen and colleagues (2016)
hypothesis (solid line) implies a transient rise in oxidative damage after the did not find changes in protein carbonyls
start of reproductive effort. In the latter two cases, the damage level is lower but rather found a negative correlation
at the end of the reproductive event than at the prereproductive stage. This in pregnant females between the num-
is not predicted in extortion for reproduction, where damage at the end of ber of fetuses in the uterus and plasma
reproduction can remain higher than damage in the non-reproductive stage. MDA values. Furthermore, plasma MDA
of adults predicted survival in a 2- to
protein carbonyl levels in the heart, liver, and follicles. This, 3-year lapse. These results suggest that females may have
however, contrasted with higher levels of the same bio- shielded offspring by reducing their own damage levels,
marker in the blood of breeding animals (Yang etal. 2013, although these results were based on blood analyses and not
Xu etal. 2014). Reduced oxidative-damage values in breed- on reproductive tissues. Moreover, we cannot fully discard
ers compared with those of nonbreeders would be required the possibility that a reduction in lipid intake or fat reserves
for shielding offspring because embryos are more sensitive during pregnancy could influence circulating lipid levels
to disturbances than adults are. The hypothesis is appealing and, accordingly, plasma MDA values (see the section on the
as it considers fitness returns derived from offspring protec- extortion-for-reproduction hypothesis below).
tion (via reproductive success).
Some weaknesses in this hypothesis can, nonetheless, Hormesis during reproduction. Intermediate levels of exposure to
be argued. First, if the mechanism is preemptive, the lower a stressor may trigger a physiological compensatory response
oxidative-damage levels found in the organs of animals dur- that could, theoretically, not only maintain fitness returns
ing the last stages of reproduction (i.e., the main empirical at suitable levels but even increase fitness. This response is
support; Yang et al. 2013, Xu et al. 2014) should also have known as hormesis (e.g., Costantini 2014). If the breeding
been present just before or at the beginning of reproduction effort is considered a mild stressor (table 1, A2), lower oxida-
(figure 4). In other words, the low value detected at the end tive-damage levels among breeders may be due to hormesis
of reproduction was interpreted as evidence of a precedent (figure 3b). However, hormesis should only be triggered
decline. However, why should we expect oxidative-damage when animals are exposed to reproductive stress (Costantini XXXX XXXX / Vol. XX No. X BioScience 7

Overview Articles

2014), whereas the preemptive response proposed in the a reproductive situation. We believe that the focus should
oxidative-shielding hypothesis should take place in advance therefore be placed on the control, nonreproducing animals.
of stress (figure 4; Blount et al. 2016). This complicates the Importantly, we assume that experimental controls showing
decision of when sampling should be performed. high oxidative damage did not suffer social stress due to iso-
In trying to explain aging (i.e., not the cost of repro- lation or stocking density. These situations lead to changes in
duction), the mitohormesis hypothesis (i.e., Tapia 2006) glucocorticoid secretion that, in turn, could increase oxida-
proposes that a longer lifespan after caloric (mostly glu- tive stress (Costantini 2016).
cose) restriction in the diet is due to a moderate increase We propose that animals prevented from breeding under
in mitochondrial ROS production. This, in turn, stimu- optimal conditions during prolonged periods would, in
lates a compensatory (hormetic) response improving long- some way, be forced to reproduce by high oxidative dam-
term oxidative stress resistance. This response would be age (figure 3c). This mechanism may also have evolved
induced by certain redox signals (reviewed in Ristow and in the wild (table 1, A3). Wild animals can experience
Schemeisser 2014). In addition to reduced glucose metabo- high resource availability, allowing breeding, but may be
lism (Schulz etal. 2007), moderately high levels of oxidative unable to mate if the population (operational) sex ratio is
damage (Tapia 2006), including lipid peroxidation, could biased. Macronutrients not diverted to reproduction can
also activate protection via mitohormesis (e.g., Ristow etal. become oxidized (Romero-Haro et al. 2015) or stimulate
2009, Labuschagne et al. 2013). In that case, the involve- NADPH oxidases (above), recreating a terminal situation
ment of mitochondria may not be strictly necessary, because (see terminal investment in Clutton-Brock 1984). This
other pro-oxidant sources (figure 2; see also Ristow and should not be limited to species assigned to the traditional
Schemeisser 2014) could trigger protection. classification of capital breeder (i.e., species mostly relying
In a reproductive context, Gago-Domnguez and col- on energy stores for reproduction; Jonsson 1997), because
leagues (2005) suggested that reduced breast cancer preva- high circulating levels of macromolecules (also present in
lence among parous women is due to the protective (perhaps income breeders) could be sufficient to trigger the mecha-
hormetic) effect of high lipid peroxidation levels during nism. High amounts of oxidized macromolecules or ROS
pregnancy. Indeed, a rise in the level of oxidative damage to would alter redox-signaling pathways, promoting hormonal
blood lipids in pregnant compared with that in nonpregnant and behavioral changes leading to reproduction with as
females has been demonstrated in humans (reviewed in Little little delay as possible (figure 3c; table 1, A1, P7). Under
and Gladen 1999) and other mammals, including rodents free-living conditions, this could be attained by dispersal
(e.g., Upreti etal. 2002, Castillo etal. 2005, but see Vitikainen to other populations. In summary, animals prevented from
etal. 2016). The same pattern has been described for protein reproducing would be spurred on by the perception that life
carbonyls (e.g., Mihu et al. 2012). The increase seems to is advancing too rapidly. This hypothesis is attractive, but is
be transient and limited to the gestation period, at least for there any supporting evidence?
lipid peroxides (Gago-Domnguez etal. 2005). However, the Cancer epidemiology suggests higher levels of oxidative
ephemeral nature of these changes does not imply positive damage in women without reproductive experience. We
effects on fitness. In fact, as far as we know, no studies have must note that cancer development is strongly linked to high
reported lower lipid peroxidation levels in mothers after birth oxidative stress levels and high activity of NADPH oxidases
compared with levels in nonreproducing women. Moreover, (Bedard and Krause 2007). Nulliparity and late age at first
if hormesis indeed acts on reproducing females, their lifes- birth in women are associated with a higher prevalence of
pans or future fecundities should be increased compared breast cancer (e.g., Ma et al. 2006), which is also the most
with those of nonbreeders (Schulz etal. 2007). Therefore, the frequent cancer in women worldwide (
hormesis hypothesis also challenges the traditional evolu- cancer-facts-figures/worldwide-data). The same association
tionary theory considering reproduction as a cost. was found in other reproductive and nonreproductive can-
cers (e.g., ovarian, brain, pancreas, colon, and liver; Martnez
Extortion for reproduction.The possibility that reproduc- etal. 1997, Merrill etal. 2005, Chang etal. 2010, Wu etal.
tion reduces oxidative damage is not as paradoxical if 2011, Schler et al. 2013). Late parturition is, nonetheless,
we consider environmental conditions and life history. related to a lower prevalence of uterine cancers, mostly cer-
Experiments preventing breeding are difficult to perform in vical cancer (Merrill etal. 2005). However, cervical cancer is
the field (e.g., it requires animal sterilization or the removal primarily due to human papillomavirus (Bosch etal. 2006).
of nest sites or mates), whereas captivity (laboratory) con- We assume that women in most demographic studies on
ditions are usually designed for optimal reproduction. In cancer have a sufficiently high availability of energy macro-
terms of life history, most laboratory studies have involved nutrients to allow for reproduction. However, we recognize
animals without reproductive experience (Speakman and that demographic studies in humans are basically correla-
Garratt 2014, Blount etal. 2016). We suggest that prevent- tions in which socioeconomic factors and other covariates
ing reproduction under optimal conditions and for more or cannot be well controlled. Therefore, causality should be
less long periods can affect the animals under study. Under inferred with caution. Moreover, we could argue that cancer
these circumstances, the organism would be better suited to is mostly detected at relatively older ages, with its impact

8 BioScience XXXX XXXX / Vol. XX No. X

Overview Articles

on individual fitness perhaps being minor. Nonetheless, young individuals that may be better able to face the chal-
fitness returns in older women may also arise via parental lenge, at least under laboratory conditions (Speakman and
care devoted to adult offspring or grandchildren (Hawkes Garratt 2014, Blount et al. 2016). In fact, we predict that the
2004). In any event, we note the lack of knowledge regarding mechanism would be reinforced when age advances (see the
the impact of cancer on wild populations of other species next section).
(Vittecoq etal. 2015).
Potential evidence may be drawn from results of an exper- Testing the hypotheses. Experiments testing reproductive costs
iment manipulating precopulatory reproductive investments are usually performed by randomly allocating animals to
in zebra finches (Romero-Haro et al. 2015). We predicted reproduction or nonreproduction groups or by manipulat-
that male zebra finches caged in view of females should show ing reproductive effort only among breeders (e.g., Alonso-
an increase in MDA levels in plasma as a result of investing Alvarez and Velando 2012, Metcalfe and Monaghan 2013).
carotenoid pigments (antioxidants) in reproduction (sexual To test the three hypotheses using these approaches, an
coloration). We found the opposite, with males housed in accurate sampling schedule is required involving repeated
view of other males showing an increase in plasma MDA measures taking into account the predicted changes in oxi-
values. Our result was, however, explained by an increase in dative damage (table 1, P1 and P2; figure 4).
the level of circulating triglycerides. Plasma MDA and even In the case of oxidative shielding, at least three sampling
hydroperoxide levels, another oxidative-damage marker events are required: two events before reproduction to docu-
(derivatives of reactive oxygen metabolities, d-ROMS; e.g., ment a preemptive decline in damage (i.e., to discard horme-
Costantini etal. 2014) are positively correlated to circulating sis) and another sampling after breeding to demonstrate that
lipid (triglycerides) levels, at least in birds (Prez-Rodrguez the potential impact was not only avoided but also reversed
etal. 2015) and humans (e.g., Toescu etal. 2004). The same (i.e., less damage among breeders; figure 4). Only noninva-
correlation is present for isoprostanes, another popular lipid sive techniques such as blood analyses, or perhaps semen
oxidation marker, in human blood (Halliwell and Lee 2011). sampling, allow for repeated measures, precluding the study
This means that recent food intake, fatness, or lipid mobi- of internal organs. The question is important if we predict
lization from stores may potentially affect these measures. that oxidative shielding should be present in those tissues
Therefore, we urge caution in direct interpretations derived that mostly influence offspring quality (i.e., the reproductive
from these parameters. Male zebra finches housed with other system; table 1, P3). Alternatively, different animals could
males showed increased body mass and plasma triglyceride be sacrificed at different phases, but this reduces the reli-
levels, probably because of lower rates of social interactions, ability of results compared with that of longitudinal analy-
and therefore less energy consumption (Romero-Haro etal. ses. In addition, experiments should include the analysis of
2015). The results, therefore, suggest that sexually mature the individual quality of descendants. Parents generating
animals with abundant energy macromolecules, circulating a stronger preemptive decline in oxidative damage should
in blood or being stored, may suffer oxidative damage when produce higher-quality descendants (i.e., lower oxidative
no mate is available (table 1, A2). damage and higher survival probabilities; table 1, P4).
In addition, recent studies testing somatic and ger- In the case of hormesis during reproduction, tissues
mline signaling networks leading to longevity extension should be sampled before and during reproduction to detect
and reduction, respectively, suggest that the first is, at least a transient rise in oxidative stress (figure 4). A third meas-
partially, attained by reorganizing lipid stores (Hansen ure should be carried out to detect the subsequent decline,
et al. 2013). Germline ablation (mostly in invertebrates) which should attain lower damage levels than those found in
favors transcription factors promoting lipase activity and control nonbreeding individuals. The positive impact of this
autophagy, lipid unsaturation, and beta-oxidation (Hansen mechanism should be tested in terms of parental longevity
etal. 2013, Lapierre etal. 2013). Accordingly, some authors but may also affect current and future reproductive success
have suggested that these signals may eliminate the excess (i.e., fitness; table 1, P4).
of those fatty acids used for reproduction and convert them Testing extortion for reproduction requires the detection
into forms that favor a long life (Ratnappan etal. 2013). This of increased levels of both macronutrients (lipids and/or
perspective agrees with the idea that high lipid accumulation proteins) and oxidized compounds (e.g., MDA or protein
for reproduction leads to higher oxidative stress in prebreed- carbonyls) among fully sexually mature and healthy individ-
ing animals. uals prevented from breeding. A decline in both parameters
Although the hypothesis may apparently reconcile empir- when reproduction is allowed to occur must also be seen
ical findings with life-history theory, one potential problem (figure 4). However, such a decrease may not result in less or
is that no study testing the oxidative cost of reproduction has equal oxidative damage in breeding individuals compared
reported (or tested) reduced longevity among nonreproduc- with that of animals in a nonreproductive stage (figure 4). In
ing captive animals showing higher oxidative-damage levels the latter two hypotheses, the type of tissue is not necessarily
(table 1, P4). This would be expected if higher oxidative related to the quality of the offspring (table 1, P4).
damage among nonbreeders leads to reduced longevity. We Experimental approaches can involve not only manipu-
must nonetheless consider that, most studies have involved lations of breeding effort but also of oxidative stress. This XXXX XXXX / Vol. XX No. X BioScience 9

Overview Articles

would allow the testing of predictions related to life history study must be improved. Importantly and not previously
traits. For instance, in the case of the hormesis-during- highlighted, it cannot be discarded that some animals
reproduction hypothesis, we should predict that reproduc- prevented from breeding may be low-quality individuals
ing animals supplied with dietary antioxidants will mount unable to reproduce. These individuals may endure greater
a weaker compensatory response and suffer a reduction in oxidative damage, buffering the statistical effect. Moreover,
fitness (shorter lives or reduced fecundity; table 1, P6). In some animals will be unable to breed quickly in the relatively
the extortion-for-reproduction hypothesis, we predict that short time period of an experimental work. In fact, experi-
fully sexually mature nonreproducing animals with ad libi- mental studies often mention that those animals allocated
tum access to macronutrients and without previous breed- to reproduction that did not breed during the study period
ing experience should delay their first reproduction when were excluded (e.g., Yang etal. 2013, Costantini etal. 2014),
antioxidant supplements are provided (table 1, P6). The last but not all the works describe this detail. The excluded
two predictions are apparently counterintuitive considering animals may never breed (i.e., may be sick individuals) or
our current knowledge on reproductive constraints (e.g., may require more time to reach a certain threshold in terms
Dowling and Simmons 2009) and the oxidative stress theory of energy, micronutrients, or signaling (hormones) allow-
of aging (Sies and Cadenas 1985). Nonetheless, they can be ing reproduction. They may also be prone to physiological
tested by correctly adjusting antioxidant doses. As alterna- corticoid-based stress under captivity conditions. Although
tives to dietary antioxidants, mitochondrial uncouplers these biases are probably subtle, they would act in opposing
and NOX inhibitors developed to combat cancer and other directions (i.e., by including and excluding the worst animals
human diseases may be promising methods to reduce ROS in nonbreeders and breeders, respectively) increasing the
generation (e.g., Salin et al. 2012, Altenhfer et al. 2015). likelihood of detecting null effects or even oxidative-damage
The simultaneous use of these two methods could allow reduction among breeders. The only solution is to test and
for testing the relative contribution of mitochondrial and control for differences between treatments in oxidative stress
extramitochondrial ROS sources to the oxidative cost of levels before the experiment. This is, however, infeasible
reproduction. when animals are sacrificed for analyzing internal tissues.
To conclude, the three hypotheses are compatible. They To demonstrate a cost in evolutionary terms, any manipu-
formulate predictions for different reproductive groups. lation of reproductive investment should have a consequence
Hormesis and oxidative shielding should act among breed- in terms of reduced fitness (e.g., Alonso-Alvarez and Velando
ers, whereas extortion for reproduction should affect 2012). This has rarely been tested for most studies testing the
nonbreeders (table 1, P1). This, in turn, should magnify OCR and is a difficult task, particularly in free-living animals
differences between groups. Nonetheless, we also predict (but see Vitikainen et al. 2016). For instance, Santos and
that hormesis during reproduction could weaken over the Nakagawa (2012) recently performed a meta-analysis review-
lifespan as we expect an aging-related decline in the capaci- ing evidence of a cost of brood or clutch size manipulations
ties for hormesis (table 1, P5; Costantini 2014). The same in wild birds, detecting weak effects on survival (only among
could apply to oxidative shielding, assuming an increased male parents experiencing increased breeding effort). An
accumulation of oxidative damage with age (i.e., Sies and absence of effective evolutionary costs would make discus-
Cadenas 1985). However, extortion for reproduction should sions about the proximate mechanisms irrelevant.
mostly act or be reinforced at advanced ages as opportunities The same argument can be applied to the choice of ana-
for breeding decline (table 1, P5). Therefore, experiments on lytical techniques. Much effort has been placed on oxidative-
comparatively young animals would be more likely to detect damage markers, because they are supposedly unambiguous.
hormesis, whereas manipulations in older individuals would However, the evidence of a connection between levels of
be more likely to support the extortion hypothesis. In sup- these parameters and longevity in longitudinal studies is still
port of the latter, studies have shown that male and female scarce. Furthermore, the search for a link to fitness should
rats prevented from reproducing throughout their lifetime not be limited to longevity but should consider future
undergo higher levels of oxidative damage compared with fecundity (e.g., Costantini et al. 2015). In fact, it has been
reproducing animals, but only at older ages (about 2 years suggested that oxidative stress may act as a reproductive
old; Alabarse etal. 2011, Da Silva etal. 2013). constraint (reviewed in Stier etal. 2012; see also Costantini
etal. 2016). If oxidative stress constrains reproduction, oxi-
Conclusions dative stress due to reproduction could also be costly by lim-
In our view, the oxidative-cost-of-reproduction hypoth- iting future reproduction. We suggest focusing on oxidative
esis does not present insurmountable theoretical problems. damage in germline DNA, which mostly promotes deleteri-
Efforts should be redirected to refine experimental designs ous mutations but may also lead to infertile adult offspring
and address new hypotheses (figure 3). The fact that virtu- (Velando etal. 2008). All parental investment in these defi-
ally all of the experiments reporting counterintuitive results cient descendants would be lost. Therefore, germline oxida-
compare reproducing versus nonreproducing animals, and tive damage might have a greater impact on parental fitness
not reproductive animals enduring different breeding effort than does oxidative damage in somatic cells. Unfortunately,
(but see Garratt et al. 2013), suggests that the first type of no work has yet addressed this question. Finally, the study of

10 BioScience XXXX XXXX / Vol. XX No. X

Overview Articles

the entire transcriptome (e.g., Rey etal. 2016) could reveal Castillo C, Hernandez J, Bravo A, Lopez-Alonso M, Pereira V, Benedito JL.
unknown signaling molecules involved in the reproductive 2005. Oxidative status during late pregnancy and early lactation in dairy
cows. Veterinary Journal 169: 286292.
trade-off, which could open the possibility of manipulations Chang CC, Chiu HF, Yang CY. 2010. Parity, age at first birth, and risk
allowing the testing of the proposed hypotheses. of death from pancreatic cancer: Evidence from a cohort in Taiwan.
Pancreas 39: 567571.
Acknowledgments Clutton-Brock TH. 1984. Reproductive effort and terminal investment in
We thank Thomas Flatt, Malene Hansen, and Hugo Aguilaniu iteropaous animals. American Naturalist 123: 212229.
for fruitful discussions on the role of reorganization of lipid Costantini D. 2014. Oxidative Stress and Hormesis in Evolutionary Ecology
and Physiology: A Marriage between Mechanistic and Evolutionary
stores and germline signals in longevity. Dr. Chen Hou also
Approaches. Springer.
contributed to clarifying the link the biosynthetic rate and . 2016. Oxidative stress as a cost of reproduction: Beyond the sim-
oxidative stress. We acknowledge Professor Neil Metcalfe plistic trade-off model. Frontiers in Ecology and Evolution 4 (art. 10).
for comments on previous versions of the text. We are also Costantini D, Casasole G, Eens M. 2014. Does reproduction protect against
grateful to Mara Benfato, Pawel Brzek, and David Costantini oxidative stress? Journal Experimental Biology 217: 42374243.
Costantini D, Goutte A, Barbraud C, Faivre B, Sorci G, Weimerskirch H,
for providing details on their experimental designs and to
Delord K, Chastel O. 2015. Demographic responses to oxidative stress
Sarah Young for her English review. We also acknowledge and inflammation in the wandering albatross (Diomedea exulans).
comments from the five reviewers that helped to improve PLOS ONE 10 (art. e0133967).
the first versions of the text. Ana Angela Romero-Haro Costantini D, Casasole G, AbdElgawad H, Asard H, Eens M. 2016.
was funded by a Formacin de Personal de Investigacin Experimental evidence that oxidative stress influences reproductive deci-
sions. Functional Ecology 30: 11691174. doi:10.1111/1365-2435.12608
grant (no. BES-2010-035013; Ministerio de Economa y
Covarrubias L, Hernndez-Garca D, Schnabel D, Salas-Vidal E, Castro-
Competitividad, MINECO, Spanish Government) and pro- Obregn S. 2008. Function of reactive oxygen species during animal
ject nos. CGL2012-40229-C02-01 and CGL2015-69338-C2- development: Passive or active? Developmental Biology 320: 111.
2-P (MINECO, Spain). Da Silva, AC, Salomon TB, Behling CS, Putti J, Hackenhaar FS, Alabarse
PVG, Schller AK, Benfato MS. 2013. Oxidative stress in the kid-
References cited ney of reproductive female rats during aging. Biogerontology 14:
Alabarse PVG, Hackenhaar FS, Medeiros TM, Mendes MFA, Viacava PR, 411422.
Schller AK, Salomon TB, Ehrenbrink G, Benfato MS. 2011. Oxidative De Jong G, van Noordwijk AJ. 1992. Acquisition and allocation of resources:
stress in the brain of reproductive male rats during aging. Experimental Genetic (co) variances, selection, and life histories. American Naturalist
Gerontology 46: 241248. 139: 749770.
Alberici LC, Oliveira HCF, Paim BA, Mantello CC, Augusto AC, Zecchin Drummond GR, Sobey CG. 2014. Endothelial NADPH oxidases: Which
KG, Gurgueira SA, Kowaltowski AJ, Vercesi AE. 2009. Mitochondrial NOX to target in vascular disease? Trends in Endocrinology and
ATP-sensitive K+) channels as signals to liver mitochondria in response Metabolism 25: 452463.
to hypertriglyceridemia. Free Radical Biology Medicine 47: 14321439. Dymkowska D, Drabarek B, Podszywaow-Bartnicka P, Szczepanowska J,
Alonso-Alvarez C, Velando A. 2012. Benefits and costs of parental care. Zabocki K. 2014. Hyperglycaemia modifies energy metabolism and
Pages 4054 in Klliker M, Smiseth PT, Royle NJ, eds. Evolution of reactive oxygen species formation in endothelial cells in vitro. Archives
Parental Care. Oxford University Press. in Biochemistry and Biophysiology 542: 713.
Alonso-Alvarez C, Bertrand S, Devevey G, Prost J, Faivre B, Sorci G. 2004. Edward DA, Chapman T. 2011. The evolution and significance of male mate
Increased susceptibility to oxidative stress as a proximate cost of repro- choice. Trends in Ecology and Evolution 26: 647654.
duction. Ecology Letters 7: 363368. Flatt T. 2011. Survival costs of reproduction in Drosophila. Experimental
Altenhfer S, Radermacher KA, Kleikers PWM, Wingler K, Schmidt Gerontology 46: 369375.
HHHW. 2015. Evolution of NADPH oxidase inhibitors: Selectivity and Flatt T, Amdam GV, Kirkwood TBL, Omholt SW. 2013. Life-history evolu-
mechanisms for target engagement. Antioxidants and Redox Signaling tion and the polyphenic regulation of somatic maintenance and sur-
23: 403427. vival. Quarterly Review of Biology 88: 185218.
Amunugama K, Jiao L, Olbricht GR, Walker C, Huang Y-W, Nam PK, Hou Forman HJ, Fukuto JM, Torres M. 2004. Redox signaling: Thiol chemistry
C. 2016. Cellular oxidative damage is more sensitive to biosynthetic rate defines which reactive oxygen and nitrogen species can act as second
than to metabolic rate: A test of the theoretical model on hornworms messengers. American Journal of Physiology: Cell Physiology 287:
(Manduca sexta larvae). Experimental Gerontology 82: 7380. C246C256.
Barja G. 2013. Updating the mitochondrial free radical theory of aging: An Gago-Dominguez M, Castelao JE, Pike MC, Sevanian A, Haile RW. 2005.
integrated view, key aspects, and confounding concepts. Antioxidants Role of lipid peroxidation in the epidemiology and prevention of
and Redox Signaling 19: 14201445. breast cancer. Cancer Epidemiology, Biomarkers and Prevention 14:
Bedard K, Krause K-H. 2007. The NOX family of ROS-generating NADPH oxi- 28292839.
dases: Physiology and pathophysiology. Physiological Reviews 87: 245313. Garratt M, McArdle F, Stockley P, Vasilaki A, Beynon RJ, Jackson MJ, Hurst
Blount JD, Vitikainen EIK, Stott I, Cant MA. 2016. Oxidative shielding and JL. 2012. Tissue-dependent changes in oxidative damage with male
the cost of reproduction. Biological Reviews 91: 483497. reproductive effort in house mice. Functional Ecology 26: 423433.
Bosch FX, Qiao Y-L, Castellsagu X. 2006. The epidemiology of human Garratt M, Pichaud N, King EDA, Brooks RC. 2013. Physiological adapta-
papillomavirus infection and its association with cervical cancer. tions to reproduction. I. Experimentally increasing litter size enhances
International Journal of Gynecology and Obstetrics 94: S8S21. aspects of antioxidant defence but does not cause oxidative damage in
Briga M, Verhulst S. 2015. What can long-lived mutants tell us about mech- mice. Journal Experimental Biology 216: 28792888.
anisms causing aging and lifespan variation in natural environments? Go YM, Jones DP. 2011. Cysteine/cystine redox signaling in cardiovascular
Experimental Gerontology 71: 2126. disease. Free Radical Biology and Medicine 50: 495509.
Brown DI, Griendling KK. 2009. NOX proteins in signal transduction. Free Halliwell B, Gutteridge JMC. 2007. Free Radicals in Biology and Medicine.
Radical Biology and Medicine 47: 12391253. Oxford University Press. XXXX XXXX / Vol. XX No. X BioScience 11

Overview Articles

Halliwell B, Lee CYJ. 2010. Using isoprostanes as biomarkers of oxidative Rey B, Dgletagne C, Bodennec J, Monternier P-A, Mortz M, Roussel D,
stress: Some rarely considered issues. Antioxidants and Redox Signaling Romestaing C, Rouanet J-L, Tornos J, Duchamp C. 2016. Hormetic
13: 145156. response triggers multifaceted anti-oxidant strategies in immature king
Hansen M, Flatt T, Aguilaniu H. 2013. Reproduction, fat metabolism, and penguins (Aptenodytes patagonicus). Free Radical Biology and Medicine
life span: What is the connection? Cell Metabolism 17: 1019. 97: 577587.
Hawkes K. 2004. The grandmother effect. Nature 428: 128129. Ristow M, Schmeisser K. 2014. Mitohormesis: Promoting health and
Hou C, Amunugama K. 2015. On the complex relationship between energy lifespan by increased levels of reactive oxygen species (ROS). Dose
expenditure and longevity: Reconciling the contradictory empirical Response 12: 288341.
results with a simple theoretical model. Mechanisms of Ageing and Ristow M, Zarse K, Oberbach A, Nora K, Birringer M, Kiehntopf M,
Development 149: 5064. Stumvoll M, Kahn CR, Blher M. 2009. Antioxidants prevent health-
Hsin H, Kenyon C. 1999. Signals from the reproductive system regulate the promoting effects of physical exercise in humans. Proceedings of the
lifespan of C. elegans. Nature 399: 362366. National Academy of Sciences 106: 86658670.
Isaksson C, Sheldon BC, Uller T. 2011. Oxidative stress into life-history Romero-Haro AA, Canelo T, Alonso-Alvarez C. 2015. Early development
biology. BioScience 61: 194202. conditions and the oxidative cost of social context in adulthood: An
Jones D. 2006. Redefining oxidative stress. Antioxidants and Redox experimental study in birds. Frontiers in Ecology and Evolution 3: 115.
Signaling 8: 18651879. Sakellariou GK, Jackson MJ, Vasilaki A. 2014. Redefining the major con-
Jonsson KI. 1997. Capital and income breeding as alternative tactics of tributors to superoxide production in contracting skeletal muscle: The
resource use in reproduction. Oikos 78: 5766. role of NAD(P)H oxidases. Free Radical Research 48: 1229.
Kirkwood TBL. 1977. Evolution of ageing. Nature 270: 301304. Salin K, Luquet E, Rey B, Roussel D, Voituron Y. 2012. Alteration of mito-
Krause K-H. 2007. Aging: A revisited theory based on free radicals gener- chondrial efficiency affects oxidative balance, development and growth
ated by NOX family NADPH oxidases. Experimental Gerontology 42: in frog (Rana temporaria) tadpoles. Journal of Experimental Biology
256262. 215: 863869.
Labuschagne CF, Stigter EC, Hendriks MMWB, Berger R, Rokach J, Salin K, Auer SK, Rudolf AM, Anderson GJ, Cairns AG, Mullen W, Hartley
Korswagen HC, Brenkman AB. 2013. Quantification of in vivo oxida- RC, Selman C, Metcalfe MB. 2015. Individuals with higher metabolic
tive damage in Caenorhabditis elegans during aging by endogenous rates have lower levels of reactive oxygen species in vivo. Biology Letters
F3-isoprostane measurement. Aging Cell 12: 214223. 11 (art. 20150538).
Lapierre LR, Silvestrini MJ, Nuez L, Ames K, Wong S, Le TT, Hansen M, Salmon AB, Marx DB, Harshman LG. 2001. A cost of reproduction in
Melendez A. 2013. Autophagy genes are required for normal lipid levels Drosophila melanogaster: Stress susceptibility. Evolution 55: 16001608.
in C. elegans. Autophagy 9: 278286. Salomon TB, Hackenhaar FS, Almeida AC, Schller AK, Gil Alabarse PV,
Little RE, Gladen BC. 1999. Levels of lipid peroxides in uncomplicated Ehrenbrink G, Benfato MS. 2013. Oxidative stress in testis of animals
pregnancy: A review of the literature. Reproductive Toxicology 13: during aging with and without reproductive activity. Experimental
347352. Gerontology 48: 940946.
Ma H, Bernstein L, Pike MC, Ursin G. 2006. Reproductive factors and Santos ESA, Nakagawa S. 2012. The costs of parental care: A meta-analysis
breast cancer risk according to joint estrogen and progesterone recep- of the trade-off between parental effort and survival in birds. Journal of
tor status: A meta-analysis of epidemiological studies. Breast Cancer Evolutionary Biology 25: 19111917.
Research 8 (art. R43). Schaar CE, Dues DJ, Spielbauer KK, Machiela E, Cooper JF, Senchuk M,
Martnez ME, Grodstein F, Giovannucci E, Colditz GA, Speizer FE, Hekimi S, Van Raamsdonk JM. 2015. Mitochondrial and cytoplas-
Hennekens C, Rosner B, Willett WC, Stampfer MJ. 1997. A prospective mic ROS have opposing effects on lifespan. PLOS Genetics 11 (art.
study of reproductive factors, oral contraceptive use, and risk of colo- e1004972).
rectal cancer. Cancer Epidemiology, Biomarkers and Prevention 6: 15. Schler S, Ponnath M, Engel J, Ortmann O. 2013. Ovarian epithelial tumors
Merrill RM, Fugal S, Novilla LB, Raphael MC. 2005. Cancer risk associated and reproductive factors: A systematic review. Archives of Gynecology
with early and late maternal age at first birth. Gynecologic Oncology and Obstetrics 287: 11871204.
96: 583593. Schulz TJ, Zarse K, Voigt A, Urban N, Birringer M, Ristow M. 2007. Glucose
Metcalfe NB, Monaghan P. 2013. Does reproduction cause oxidative stress? restriction extends Caenorhabditis elegans lifespan by inducing mito-
An open question. Trends in Ecology and Evolution 28: 347350. chondrial respiration and increasing oxidative stress. Cell Metabolism
Mihu D, Sabu L, Costin N, Ciortea R, Mluan A, Mihu CM. 2012. 6: 280293.
Implications of maternal systemic oxidative stress in normal pregnancy Seah NE, de Magalhaes Filho CD, Petrashen AP, Henderson HR, Laguer
and in pregnancy complicated by preeclampsia. Journal Maternal and J, Gonzalez J, Dillin A, Hansen M, Lapierre LR. 2016. Autophagy-
Neonatal Medicine 25: 944951. mediated longevity is modulated by lipoprotein biogenesis. Autophagy
Mller AP, Biard C, Blount JD, Houston DC, Ninni P, Saino N, Surai 12: 261272.
PF. 2000. Carotenoid-dependent signals: Indicators of foraging effi- Seehuus S-C, Norberg K, Gimsa U, Krekling T, Amdam GV. 2006.
ciency, immunocompetence or detoxification ability? Avian and Poultry Reproductive protein protects functionally sterile honey bee workers
Biology Reviews 11: 137159. from oxidative stress. Proceedings of the National Academy of Sciences
Olsson M, Wilson, M, Uller T, Mott B, Isaksson, C. 2009. Variation in levels 103: 962967.
of reactive oxygen species is explained by maternal identity, sex and Serpillon S, Floyd BC, Gupte RS, George S, Kozicky M, Neito V, Recchia
body-size-corrected clutch size in a lizard. Naturwissenschaften 96: F, Stanley W, Wolin MS, Gupte SA. 2009. Superoxide production
2529. by NAD(P)H oxidase and mitochondria is increased in genetically
Prez-Rodrguez L, Romero-Haro AA, Sternalski A, Muriel J, Mougeot F, obese and hyperglycemic rat heart and aorta before the development
Gil D, Alonso-Alvarez C. 2015. Measuring oxidative stress: The con- of cardiac dysfunction: The role of glucose-6-phosphate dehydroge-
founding effect of lipid concentration in measures of lipid peroxidation. nase-derived NADPH. American Journal of Physiology: Heart and
Physiological and Biochemical Zoology 88: 345351. Circulatory Physiology 297: H153H162.
Ratnappan R, Amrit FRG, Chen S-W, Gill H, Holden K, Ward J, Yamamoto Sies H, Cadenas E. 1985. Oxidative stress: Damage to intact cells and organs.
KR, Olsen CP, Ghazi A. 2014. Germline signals deploy NHR-49 to Philosophical Transactions of the Royal Society B 311: 617631.
modulate fatty-acid -oxidation and desaturation in somatic tissues of Speakman JR. 2008. The physiological costs of reproduction in small mam-
C. elegans. PLOS Genetics 10 (art. e1004829). mals. Philosophical Transactions of the Royal Society B 363: 375398.
Ren Y, Hughes KA. 2014. Vitellogenin family gene expression does not Speakman JR, Garratt M. 2014. Oxidative stress as a cost of reproduction:
increase Drosophila lifespan or fecundity. F1000Research 3 (art. 125). Beyond the simplistic trade-off model. BioEssays 36: 93106.

12 BioScience XXXX XXXX / Vol. XX No. X

Overview Articles

Stier A, Reichert S, Massemin S, Bize P, Criscuolo F. 2012. Constraint Wang Y, Salmon AB, Harshman LG. 2001. A cost of reproduction: Oxidative
and cost of oxidative stress on reproduction: Correlative evidence in stress susceptibility is associated with increased egg production in
laboratory mice and review of the literature. Frontiers in Zoology 9 Drosophila melanogaster. Experimental Gerontology 36: 13491359.
(art. 37). Wiersma P, Selman C, Speakman JR, Verhulst S. 2004. Birds sacrifice oxida-
Tapia PC. 2006. Sublethal mitochondrial stress with an attendant stoi- tive protection for reproduction. Biology Letters 271: S360S363.
chiometric augmentation of reactive oxygen species may precipitate William GC. 1957. Pleiotropy, natural selection, and the evolution of senes-
many of the beneficial alterations in cellular physiology produced by cence. Evolution 11: 398411.
caloric restriction, intermittent fasting, exercise and dietary phytonu- Wu C-H, Chan T-F, Changchien CC, Yang CY. 2011. Parity, age at first birth,
trients: Mitohormesis for health and vitality. Medical Hypotheses 66: and risk of death from liver cancer: Evidence from a cohort in Taiwan.
832843. Journal of Gastroenterology and Hepatology 26: 334339.
Tatar M. 2001. Senescence. Pages 128141 in Fox CW, Roff DA, Fairbaim Xu YC, Yang DB, Speakman JR, Wang DH. 2014. Oxidative stress in
DJ. Evolutionary Ecology: Concepts and Case Studies. Oxford University response to natural and experimentally elevated reproductive effort is
Press. tissue dependent. Functional Ecology 28: 402410.
Tatar M, Carey JR. 1995. Nutrition mediates reproductive trade-offs with Yang D-B, Xu Y-C, Wang D-H, Speakman JR. 2013. Effects of reproduction
age-specific mortality in the beetle Callosobruchus maculatus. Ecology on immuno-suppression and oxidative damage, and hence support or
76: 20662073. otherwise for their roles as mechanisms underpinning life history trade-
Toescu V, Nuttall SL, Martin U, Nightingale P, Kendall MJ, Brydon P, Dunne offs, are tissue and assay dependent. Journal of Experimental Biology
F. 2004. Changes in plasma lipids and markers of oxidative stress in 216: 42424250.
normal pregnancy and pregnancies complicated by diabetes. Clinical Zera AJ, Harshman LG. 2001. The physiology of life history trade-offs in
Science 106: 9398. animals. Annual Reviews in Ecology and Systematics 32: 95126.
Upreti K, Chaki SP, Misro MM. 2002. Evaluation of peroxidative stress and
enzymatic antioxidant activity in liver and kidney during pregnancy Carlos Alonso-Alvarez ( is affiliated with the
and lactation in rats. Health and Population: Perspectives and Issues Departamento de Ecologa Evolutiva at the Museo Nacional de Ciencias
25: 177185. Naturales (National Museum of Natural History) which is part of the Consejo
Velando A, Torres R, Alonso-Alvarez C. 2008. Avoiding bad genes: Superior de Investigaciones Cientficas (CSIC; the Spanish National Research
Oxidatively damaged DNA in germ line and mate choice. Bioessays 30: Council), in Madrid, Spain. Tara Canelo and Ana ngela Romero-Haro are
12121219. affiliated with the Instituto de Investigacin en Recursos Cinegticos (IREC;
Vittecoq M, etal. 2015. Animal behaviour and cancer. Animal Behaviour Institute for Game and Wildlife Research), in Ciudad Real, Spain, which is
101: 1926. also part of CSIC and University of Castilla La Mancha (UCLM). XXXX XXXX / Vol. XX No. X BioScience 13